Skill or luck · 60 seconds

1,000 monkeys picked trades at random. One of them looks like a genius.

Can you spot the real strategy? Three rounds, one sealed vault, no sign-up.

A cloud of random monkeys' equity curves. The luckiest one rises above the rest and looks like skill; when the sealed final 30% is opened, the engine's honest grade for it is shown.
1,000 random monkeys

How it works

Every backtest is the best of all the versions you tried.

  1. Count every try

    Tune a strategy 200 times and the best version will look brilliant, even on pure noise. The engine raises the bar for every version you tried.

  2. Seal the last part

    The final 30% of your data goes into a vault the tuning never saw. Luck usually runs out in there; skill tends to keep its slope.

  3. Race random strategies

    Thousands of no-skill strategies run on data shaped like yours. If they match your result often, the grade says so: letter, word and colour.

Why you can check our work

No opinions. One open engine.

Your file stays on your device

Quick Check runs the engine inside your browser. Nothing you upload is sent anywhere or stored.

Same file, same grade

Grades come from the open-source Embargo engine. It is deterministic: run it yourself with one pip command and get the same report.

Every number has a source

The game and the toy replay real engine runs from published seeds. The site never invents a grade or a statistic.

Plain words, no tips

A grade is a method check on the file you uploaded. It never tells you what to trade and says nothing about the future.

Think your backtest is good?
Count your tries.